3 considerations before moving to the cloud

Despite the urgency of IT modernization and the federal government's cloud-first mandate, many agencies unfortunately still find themselves lagging when it comes to cloud adoption. While cloud migration is a massive endeavor, it doesn’t have to be unmanageable, let alone impossible. Indeed, much of the hassle and potential complications can be mitigated, if not altogether avoided, by developing a comprehensive strategy for migrating mission-critical applications to the cloud. Here are three recommended best practices agencies should consider when developing their cloud migration strategy.

1. Assessing current workloads

It would be nice if agencies could simply lift and shift their applications to the cloud, but in the majority of cases, this will not be practical. There are multiple factors to consider, from priorities and objectives, to timing, resources, architectural considerations and more. These competing factors must be addressed and triaged as agencies build out their strategy.

CTOs should start by getting a complete understanding of the applications and workloads currently running on premises, including backups and storage that are too easily overlooked. When planning a move to the cloud, it’s critical that IT managers review the entire application stack, as well as the supporting IT processes and operations so they can identify and prioritize the workloads that may be an immediate fit for the cloud and which ones may require considerable refactoring before migration.

Remember, the cloud provider supplies a blank canvas, and the bulk of the work comes down to how the client architects the solution and configures the new cloud environment. CTOs may think they know enough about their systems, but once they start digging, they’re guaranteed to find pieces of the puzzle they missed.

2. Prioritizing applications

This is the heart and soul of the migration process. IT teams must analyze existing applications to determine how they are constructed. Because all applications are different, some may be efficient and well-structured, while others are bulky, monolithic and full of complex business processes.

Once all applications have been thoroughly reviewed, the prioritization can begin. Some will be easy to migrate, especially if they’re already virtualized and can be moved with less rework. These applications could benefit from being componentized or containerized, thus making them more manageable and portable in the future.

Other applications may need to be rewritten, restructured or, for the impossibly immovable, left untouched where they are. Part of prioritization is determining what can or should be left behind. Luckily, if the workload has been properly assessed, this part of the process will be much easier.

3. Application optimization

Optimization doesn’t start after applications and workloads have been migrated to the cloud. The real optimization work starts on premise, before any of the applications and workloads have been moved. Once everything moves to the new cloud platform, it will be much more difficult to troubleshoot and identify application and performance issues unless a baseline for optimal performance was established prior.

Creating a meaningful application performance baseline before the move means that IT teams must assess the performance of every component the application depends on, including database performance, queries hitting the database, the application server, response time to the user and everything in between.

With the application baselines established, CTOs will have a much easier time pinpointing performance issues and bottlenecks after the applications have been migrated to the cloud. Maybe the query to the database should be rewritten for the new environment. Maybe some of the applications have to be broken up or rewritten. Whatever the case, by establishing baselines at the beginning of the process, troubleshooting issues with application performance becomes a more productive task.

A blank canvas

At the end of the day, cloud providers are essentially supplying commodities, offering similar suites of services and trying to differentiate themselves through offerings such as artificial intelligence. The onus is on the agency to reconstruct and manage the application stack in a way that’s custom tailored to their specific needs, and there is no shortcut. The only way to guarantee that migration leads to improved performance is to take the time to analyze the workload, establish baselines and prioritize which applications can be moved, how they should be moved and what components will be left behind. By doing so, agencies will set themselves up with documented, inventoried and prioritized apps and an easy to follow roadmap to cloud migration success.